272 Florida Entomologist 84(2) June 2001

PUPATION BIOLOGY OF FRANKLINOTHRIPS ORIZABENSIS (THYSANOPTERA: ) AND HARVESTING AND SHIPPING OF THIS PREDATOR

MARK S. HODDLE, KIKUO OISHI, AND DAVID MORGAN Department of Entomology, University of California, Riverside, CA 92521, USA

ABSTRACT In the laboratory, 82-99% of late second instar Franklinothrips orizabensis Johansen larvae abandoned branches and artificial branches constructed of wooden dowels and were recovered below branches trapped on tangle foot coated plastic sheets, suggesting a prefer- ence by this life stage for selection of pupation sites beneath host plants. Of three media tested (coarse and fine vermiculite, and parafilm cones) for harvesting F. orizabensis pupae in cocoons, parafilm cones were most easily harvestable from colonies, and 44% of deployed late second stage larvae that were recovered used parafilm cones for pupation in experimental cages. Harvesting and shipping trials using aspirated adult F. orizabensis or pupae in para- film cones showed significant differences in survivorship when held in the laboratory or shipped round trip from Riverside, California to Amherst, Massachusetts. Survivorship of aspirated adults was reduced on average by 41% following shipping, and mortality was high- est for adult males. Transit survivorship was increased by 53% if F. orizabensis were shipped as cocoons in parafilm cones. Inclusion of ice packs in polystyrene boxes did not significantly increase survivorship rates for F. orizabensis adults or pupae that were either retained in the laboratory or shipped. This result may have been an artifact resulting from the time of year (i.e., May and temperatures were moderately cool) when shipping trials were conducted.

Key Words: Franklinothrips, pupation biology, harvesting, shipping

RESUMEN En el laboratorio, 82-99% de las larvas en el segundo instar de Franklinothrips orizabensis abandonaron las ramas de aguacate y las ramas artificiales construidas de clavijas de ma- dera y fueron recuperadas bajo ramas atrapadas en hojas plásticas con capa pegajosa “tan- gle foot”, sugiriendo una preferencia de selección de lugares de pupación debajo de plantas huéspedes por esta etapa de vida. De les tres medios probados (vermiculita fina y ordinaria, y conos de parafina) para cosechar pupas de Franklinothrips orizabensis en capullos, conos de parafina fueron los mas fáciles de cosechar de las colonias, y el 44% de larvas de segunda etapa que fueron recuperadas usaron conos de parafina para pupación en jaulas experimen- tales. Ensayos de cosecha y embarque usando adultos aspirados de Franklinothrips oriza- bensis o pupas en conos de parafina demostraron diferencias significativas en supervivencia al ser llevadas a cabo en el laboratorio o embarcadas ida y vuelta desde Riverside, California hasta Amherst, Massachussets. Supervivencia de adultos aspirados fue reducida un prome- dio de 41% después de embarque, y la mortalidad fue mayor entre machos adultos. Supervi- vencia de transporte fue incrementada por 53% si Franklinothrips orizabensis eran embarcadas como capullos en conos de parafina. La inclusión de bolsas de hielo en cajas de poliestireno no incrementó significativamente las cantidades de supervivencia para adultos o pupas de Franklinothrips orizabensis que fueron retenidas en el laboratorio o embarcadas. Este resultado puede haber sido un artefacto resultando por el tiempo del año (es decir, Mayo y las temperaturas eran moderadamente frías) cuando los ensayos de embarque fueron lle- vados a cabo.

Inoculative, augmentative, and inundative bi- facilitate rapid pest outbreaks as natural ological control programs targeting pestiferous enemy densities have declined because of lack of thrips in perennial outdoor crops have seldom prey. (3) Incompatibility of thrips natural ene- been successful (Parker & Skinner 1997). The ma- mies with broad-spectrum insecticides used to jor limiting factors that have been identified as control other crop pests. (4) The high cost of insec- constraints on successful thrips biological control tary-reared natural enemies make large-scale programs for outdoor crops are: (1) lack of effec- field releases for thrips control into outdoor crops tive resident natural enemies that respond in a of low value unfeasible economically (Grafton- rapid density dependent manner to increasing Cardwell & Ouyang 1995a; Parker & Skinner thrips densities. (2) Thrips phenology and life cy- 1997; Parrella & Lewis 1997). Natural enemies cle characteristics can result in long periods of that have been commonly observed with phytoph- low thrips densities. In some instances this may agous thrips in perennial tree crops are phy-

Hoddle et al.: Franklinothrips pupation biology 273

toseiid mites and predatory thrips (Parker & ments and diets for mass rearing this predator Skinner 1997, Parrella & Lewis 1997). have been evaluated (Hoddle et al. 2000, 2001) In some instances, thrips biological control by and limited field trials evaluating releases of Type IV phytoseiid mites (i.e., specialized pollen F. orizabensis against S. perseae on have feeders that exhibit generalist predatory activity been conducted (Silvers 2000). [McMurtry & Croft 1997]) can be enhanced by Augmentative field releases of mass-reared F. pollen bearing wind break trees (Grout & Rich- orizabensis onto avocado trees failed to signifi- ards 1990, 1992a,b) or cover crops in orchards cantly suppress S. perseae populations in Califor- (Grafton-Cardwell et al. 1999a). Crop manage- nia. This may have been due more to the poor ment practices such as pruning that enhances quality of adults which suffered high mortality succulent leaf material can also promote in- (>50%) after shipping from an insectary in Europe creased densities of Type IV phytoseiids (Grafton- rather than the inherent ineffectiveness of Cardwell & Ouyang 1995b; Grafton-Cardwell F. orizabensis (Silvers, 2000). In order to fully ex- 1997). Augmentative releases of insectary-reared plore the potential of augmentative releases of phytoseiids for thrips control have been shown to F. orizabensis for controlling of S. perseae in Cali- significantly reduce densities of Scirtothrips citri fornia avocado orchards, low impact techniques for Moulton (Thysanoptera: Thripidae) on citrus (Cit- harvesting and distribution that minimize transit rus spp.), but, at present, this technology is not mortality of this predator are required. One possi- cost effective (Grafton-Cardwell & Ouyang ble approach would be to collect and transport F. 1995a; Grafton-Cardwell et al. 1999b). orizabensis when larvae are pupating within pro- Scirtothrips perseae Nakahara (Thysanoptera: tective silk cocoons as opposed to aspirating and Thripidae) was first discovered damaging avocado shipping adults. The purpose of this study was to foliage and fruit in southern California orchards investigate the pupation biology and behavior of F. in 1996, and at time of discovery was a species orizabensis, and to develop techniques based on an new to science (Nakahara 1997). In 1998, crop understanding of pupation behavior for collecting losses due to down-graded fruit and increased and shipping pupating larvae and compare survi- production costs due to S. perseae feeding damage vorship rates to currently employed methods for were estimated to have cost California growers collecting and shipping adult predators. (US) $7-$13 million (Hoddle et al. 1998, 1999). Foreign exploration efforts to determine the na- MATERIALS AND METHODS tive range of S. perseae indicate that this pest is of Central American origin (Hoddle et al. 1999) and Franklinothrips orizabensis Colony associated natural enemies collected concurrently from avocados in Latin America have included Franklinothrips orizabensis colonies were Franklinothrips spp. (Hoddle, unpublished). maintained in cages in a temperature controlled Surveys in southern California avocado or- room (25°C, 60% RH, L:D 14:10) on lima beans chards for indigenous natural enemies associated (Phaseolus lunatus Linnaeus variety “Baby Ford- with S. perseae have revealed that an undescribed hook” [plants are needed for oviposition of eggs by species of Franklinothrips orizabensis Johanson female F. orizabensis]) at the University of Cali- (Thysanoptera: Aeolothripidae) is the dominant fornia Riverside, California, USA. Colony reared predator where this pest has attained high densi- F. orizabensis were fed irradiated Ephestia kueh- ties (>10 S. perseae larvae per leaf) (Hoddle, un- niella (Zeller) (Lepidoptera: Pyralidae) eggs (sup- published). Female Franklinothrips spp. lay eggs plied by Beneficial Insectaries, Oak Run, CA, directly into plant tissue. Developing larvae pass USA) which were liberally deposited on upper through two instars (second instars are distin- surfaces of horizontal bean leaves. Our colony guished from firsts by red hypodermal pigments), was initiated with adult F. orizabensis collected before pupating within silk cocoons which are from an avocado orchard infested with S. perseae spun from secretions produced from the anal re- in Fallbrook, California, USA. Adult progeny pro- gion (Arakaki & Okajima 1998). duced by field collected F. orizabensis used to ini- Unlike Euseius tularensis Congdon (Acari: tiate the colony were deposited with the Phytoseiidae), which in some instances can regu- Systematic Entomology Laboratory, USDA-ARS, late citrus thrips, S. citri (Moulton) (Grafton- Beltsville, Maryland, USA and identified as Cardwell & Ouyang 1995b), E. hibisci (Chant) a F. orizabensis by Dr. S. Nakahara. common Type IV phytoseiid in southern Califor- nia avocado orchards, has not been observed to re- Source of Second Instar Franklinothrips orizabensis spond in a significant density dependent manner Larvae for Experiments to increasing S. perseae populations (Hoddle un- published). Because of industry interest, the po- Undersides of mature ‘Hass’ avocado leaves tential for using F. orizabensis for augmentative collected from the Biological Control Grove (F. releases against S. perseae in avocado orchards is orizabensis has not been collected here) at the being investigated. Optimal temperature require- University of California, Riverside were pre-

274 Florida Entomologist 84(2) June 2001

sented to colony-reared female F. orizabensis for clamp. A second versatile clamp was used to hold oviposition (Hoddle et al. 2000). Adult females a clear sheet of plastic (100cm × 30 cm) covered were confined with male F. orizabensis within with tanglefoot under the artificial branch. Prox- modified Munger cells (Munger 1942, Morse et al. imal ends of artificial branches were coated with 1986), fed irradiated E. kuehniella eggs, and left tanglefoot to prevent larval escape. Outlines of to oviposit in temperature cabinets (25°C, L:D artificial branches were drawn in tanglefoot. Two 14:10 h). types of artificial branches were used: (1) wooden At the end of the 24 h oviposition period, the leaf dowels without pupation refugia, and (2) wooden area enclosed by the Munger cell which was ex- dowels with pupation refugia. Refugia consisted posed to ovipositing females was excised from the of four 3mm diameter holes drilled equidistant avocado leaf. Trimmed leaves were labeled, placed around the dowel’s circumference at 20 cm inter- on water-saturated foam pads in stainless steel vals from the distal end. Artificial pupation sites trays, and incubated at 25°C. Leaves were exam- were provided to determine the propensity of ined daily for emerged F. orizabensis larvae. F. orizabensis larvae to stay on branches to pu- Emerged first instar larvae were collected with a pate when abundant sites were available in com- fine camel hair brush, placed individually in 1 dram parison to artificial branches without refugia, or glass shell vials with irradiated E. kuehniella eggs naturally occurring pupation sites on avocado as food, and vials were sealed with parafilm® branches. Ten late second instar F. orizabensis (American National Can, WI). Larvae were reared larvae of known age were placed on avocado leaf in temperature controlled cabinets (25°C, L:D 14:10 bouquets attached to artificial branches and sup- h) in shell vials until second instars which were plied with irradiated E. kuehniella eggs as food. then used for experiments. Late second instar lar- Larvae were left for three days and numbers of vae select pupation sites and spin pupation cocoons larvae on tanglefoot covered plastic sheets were within which propupal and pupal stages occur. recorded. The perpendicular distance jumped from branches was determined by measuring dis- Pupation Behavior tances from larvae to drawn branch outlines in tanglefoot. Artificial branches were examined un- Pupation behavior experiments were con- der a dissecting microscope and numbers of lar- ducted in the laboratory at 25°C and ambient vae pupating on wooden dowels were recorded. light to determine the proportion of F. orizabensis Experiments were replicated 10 times for artifi- larvae that pupate on avocado branches and arti- cial branches with and without refugia. ficial wooden branches. Pupation on avocado branches. Individual Pupal Harvesting Techniques branches were selected on small (1.5-2 m tall) Hass avocado trees planted in 20 liter plastic pots. The pupation behavior study outlined above in- Tanglefoot was applied at the base of branches at dicated that >80% of late second instar F. oriza- trunk attachment to prevent larvae leaving ex- bensis larvae would fall or jump from natural and perimental branches. Clear plastic sheets (60 cm artificial branches to seek pupation sites below × 40 cm) covered with tanglefoot were placed un- branches (see Results Section). Since late second der each experimental branch and the outline of instar larvae spin silk cocoons we sought to iden- the overhanging branch was drawn in the tangle- tify a media within which larvae would pupate foot. Ten late second instar F. orizabensis larvae of that could allow easy harvesting of F. orizabensis known age were placed on avocado leaves and in cocoons. Evaluations of pupation media were supplied with irradiated E. kuehniella eggs as conducted in temperature controlled cabinets food. Larvae were left for three days and then (25°C, L:D 14:10). Five late second instar F. oriza- numbers recovered from tanglefoot covered plas- bensis larvae were placed with pupation media in tic sheets were recorded, and the perpendicular glass Petri dishes (9.5 cm diameter) with lids distance jumped from branches was determined which were sealed with parafilm and left for three by measuring distances from larvae to drawn days in temperature controlled cabinets. Cocoons branch outlines in tanglefoot. Avocado branches were then located in Petri dishes and numbers of were removed from trees and examined under a adult F. orizabensis that successfully emerged fol- dissecting microscope and numbers of larvae pu- lowing harvesting were recorded daily. pating on branches were recorded. This Vermiculite. Vermiculite (Therm-o-Rock West experiment was replicated 10 times. Inc., Chandler, AZ) was sieved into two grades Pupation on artificial branches. Artificial coarse (all granules that were not retained by 20 branches were constructed of wooden dowels mesh USA Standard Testing Sieve) and fine (all 90cm long and 0.75 cm wide and bouquets of avo- granules not retained by 35 mesh USA Standard cado leaves wrapped in moist paper towels were Testing Sieve). Five ml of each material was attached to distal ends of dowels with parafilm. placed into glass Petri dishes and the ability to Proximal ends of artificial branches were at- harvest cocoons spun onto vermiculite granules tached to a 1.2 m ringstand with a versatile was assessed after three days. Each vermiculite Hoddle et al.: Franklinothrips pupation biology 275 treatment was replicated 10 times. Collection of Harvesting of Franklinothrips orizabensis Pupae cocoons was attempted by sieving vermiculite from Laboratory Colonies with a 20 mesh sieve based on the assumption that cocoons encased with vermiculite would be Parafilm cones were readily used as pupation too large to pass through and would be retained. sites by F. orizabensis larvae in glass Petri dishes Parafilm cones. Parafilm cones (diameter 3 mm, and small rearing cages (see Results Section). The height 5 mm) (Fig. 1) were constructed by wrap- efficacy of using parafilm cones to harvest F. oriz- ping strips of parafilm (4 mm wide, 15 mm long) abensis pupae was determined in laboratory colo- around the tapered end of a camel hair brush. nies. F. orizabensis colonies used for this Five parafilm cones were placed in each Petri experiment were fed irradiated E. kuehniella dish, and this treatment was replicated 10 times. eggs and maintained in cages (75 cm × 40 cm × 45 Cone utilization for pupation by F. orizabensis cm) with 9-12 lima bean plants. Plants were larvae was visually assessed after three days. raised 3 cm from the cage floor on metal pipes Parafilm cones and fine vermiculite. Five para- fixed to cage walls. Under these metal supports a film cones and 5 mls of fine vermiculite were com- clear plastic sheet (70 cm × 30 cm) was positioned bined in a glass Petri dish. This treatment was holding 100 parafilm cones (Fig. 2). Parafilm replicated 6 times, and utilization of either cones cones were left beneath plants for three days be- or vermiculite for pupation was assessed after fore the plastic sheet with cones was removed. three days. Cones containing pupae were counted, held in Petri dishes in a temperature controlled cabinet ° Selection of Pupation Sites in Small Rearing Cages (25 C, L:D 14:10), and number and sex of emerg- ing adult F. orizabensis was recorded daily. This Thirty individually-reared late second instar trial was replicated 10 times and approximately F. orizabensis were hand placed on four lima bean 300 adult F. orizabensis were present in mass- plants contained within rearing cages (30 cm × 30 rearing cages each time. cm × 37.5 cm), fed irradiated E. kuehniella eggs which were liberally deposited on upper surfaces Shipping Adult and Pupal Franklinothrips orizabensis of horizontal bean leaves, and maintained in a temperature controlled room (25°C, 60% RH, L:D Thirty individually-reared adult F. orizabensis 14:10) for three days. Cages were supplied with (15 male and 15 female; all were 1-2 days of age) 50 parafilm cones which were distributed on pot- were aspirated into three dram glass shell vials, ting media and cage floors. After three days, num- supplied with irradiated E. kuehniella eggs as bers of F. orizabensis cocoons in parafilm cones, in food, and sealed with a wad of cotton wool. Thirty cage corners and sleeves, on bean plants, and in F. orizabensis pupae (1-2 days of age) in parafilm vermiculite in which bean plants were growing cones were placed in three dram shell vials and (plants were destructively sampled) were deter- sealed with cotton wool. Aspirated adults and pu- mined. After inspection for cocoons, experimental pae in parafilm cones were each subjected to one cages were examined every two days for 10 days of the following treatments: for adults that emerged from cocoons that were Aspirated adults and cocoons retained in the not detected during initial cage inspection. This laboratory. The purpose of this experiment was to treatment was replicated 10 times. quantify predator mortality in the absence of shipping stress. Aspirated adults and pupae in parafilm cones were placed in sealed polystyrene foam boxes (20.5 cm × 16.5 cm × 15.5 cm) with or without ice packs. Temperatures and humidities in boxes were recorded every 10 mins with Hobo data loggers (Onset Computer Corp., Pocasset, MA). Survivorship of aspirated adults was re- corded after 48 h and pupae in parafilm cones were removed from boxes after 48 h and reared in a temperature controlled cabinet (25°C, L:D 14:10) to determine emergence rates and sex. This treatment was replicated five times. Aspirated adults and cocoons shipped to Mas- sachusetts. The purpose of this experiment was to quantify predator mortality due to shipping stress. Adult and pupal F. orizabensis were col- lected, prepared, and shipped in polystyrene foam boxes with or without ice-packs with Hobo data Fig. 1. Schematic of a parafilm cone used for harvest- loggers as described for the laboratory retention ing Franklinothrips orizabensis pupae. study. Boxes were shipped Federal Express prior- 276 Florida Entomologist 84(2) June 2001

Fig. 2. Parafilm cones on a clear plastic sheet that is placed beneath elevated bean plants in a Franklinothrips oriz- abensis rearing cage. Franklinothrips orizabensis pupae are harvested from the laboratory colony in parafilm cones. ity overnight to Amherst, Massachusetts and star F. orizabensis larvae were recovered (Table 1). then returned immediately the next day by Fed- On avocado branches and artificial branches with eral Express priority overnight to Riverside, Cal- refugia <20% of recovered larvae pupated on ifornia. Upon receipt in Riverside, proportion of branches, and the majority were recovered from adults surviving was determined, and pupae in tanglefoot coated plastic below branches. This re- parafilm cones were removed from boxes and sult indicated that late second instar F. orizaben- reared in a temperature controlled cabinet (25°C, sis larvae may prefer to pupate below trees, or L:D 14:10) to determine emergence rates and sex. suitable sites for pupation on branches were ei- Round trip transit time from California to River- ther not available or located. Artificial branches side was approximately 48 h. This treatment was without pupation refugia resulted in 100% branch replicated five times. abandonment by larvae (Table 1). Of larvae recov- Analysis of collecting and shipping data. Anal- ered from plastic sheets, 80-98% were found ysis of adult and cocoon survivorship data was within 0-2.5 cm of branches, indicating that larvae performed on logit transformed data (ln live/ fall rather than actively jump away from branches dead) after weighing for sample size. The effects when searching for pupation sites below host of shipping, cooling, life stage, and sex on survi- plants (Table 1). High rates of branch abandon- vorship of F. orizabensis were tested for signifi- ment may make it possible to collect F. orizabensis cance using Chi-square analysis and pair-wise T- below host plants if suitable pupation sites for co- test comparisons (0.025 level of significance). coon construction can be provided that allow for Analysis of temperature and humidity data. easy harvesting of this life stage. Temperature and humidity recorded for each day of the two day period for both laboratory held and Pupal Harvesting Techniques shipped polystyrene foam boxes was analyzed us- ing ANOVA and Tukey’s Studentized range test In coarse vermiculite, 80% of deployed late sec- for means separation (0.05 level of significance). ond instar F. orizabensis larvae pupated success- fully (Table 2). Use of this media for harvesting RESULTS cocoons was unsuitable as larvae spun silk co- coons between vermiculite particles and the bot- Pupation Behavior toms and sides of Petri dishes, or crawled into cracks in vermiculite particles making detection In all three treatments (i.e., natural avocado difficult. Consequently, a technique for harvest- branches and artificial branches with and without ing pupal F. orizabensis based on sieving for pupation refugia) >90% of deployed late second in- cocoons encased in vermiculite would not be prac- Hoddle et al.: Franklinothrips pupation biology 277

TABLE 1. MEAN PERCENTAGE RECOVERY (± SE) OF FRANKLINOTHRIPS ORIZABENSIS LARVAE RELEASED ONTO AVOCADO BRANCHES, ARTIFICIAL BRANCHES LACKING PUPATION SITES, AND ARTIFICIAL BRANCHES WITH PUPATION SITES.

Artificial branches Artificial without branches with Avocado pupation pupation branches refugia refugia

% Total larval recovery 93.00 ± 2.13 98.00 ± 1.33 95.00 ± 1.67 % Larvae pupating on branch 1.11 ± 0.01 0.00 ± 0.00 17.51 ± 0.01 % Larvae recovered from sticky traps 0-2.5cm from branch 97.78 ± 0.01 90.89 ± 0.01 80.38 ± 0.01 % Larvae recovered from sticky traps 2.5-5.0 cm from branch 1.11 ± 0.01 9.11 ± 0.07 2.11 ± 0.01 % Larvae recovered from sticky traps >5.0 cm from branch 0.00 ± 0.00 0.00 ± 0.00 0.00 ± 0.00 tical. The use of fine vermiculite resulted in 100% vae (Table 4). Greater numbers of females were mortality of larvae as this grade probably acted as harvested using parafilm cones, and 0-67% (mean an abrasive desiccant (Table 2). 35% ± 6.50 [SE]) of cones would have more than In Petri dishes, parafilm cones were utilized as one pupating larva. Adult emergence rates from pupation sites by 80% of deployed late second in- parafilm cones placed in colony cages was high, star F. orizabensis larvae of which 72% emerged >97% (Table 4). successfully. All recovered larvae had spun co- coons inside parafilm cones (Table 2). Parafilm Shipping Adult and Pupal Franklinothrips orizabensis cones combined with fine vermiculite resulted in 67% successful pupation and all recovered pupae Handling and shipping. Survivorship of adult were found inside parafilm cones. Provision of F. orizabensis was significantly affected by ship- parafilm cocoons mitigated the adverse effects of ping (χ2 = 44.39, df = 1, p < 0.0005) with survivor- fine vermiculite on pupation success rates (Table ship rates of adults being reduced on average by 2). Parafilm cones were the most successful media 41% in comparison to F. orizabensis that were tested for harvesting F. orizabensis pupae. held in the laboratory and not subjected to ship- ping stress (Fig. 3). Adult male F. orizabensis suf- Selection of Pupation Sites in Small Rearing Cages fered significantly higher mortality rates than females (χ2 = 34.18, df = 1, p < 0.0005) with sur- Of the 300 second instar F. orizabensis larvae vival being reduced on average by 24% (Fig. 3). released into small rearing cages 58% were recov- Survivorship of F. orizabensis was significantly ered. The highest numbers of recovered pupae increased when predators were shipped as co- were found inside parafilm cones (Table 3). No pu- coons (χ2 = 185.10, df = 1, p < 0.0005) and transit pating larvae were found on bean leaves or stems. survivorship was increased on average by 53%. Approximately 10% of pupae were found inside (Fig. 3). Inclusion of icepacks in polystyrene foam seed coats attached to cotyledons, and a similar boxes that were either retained in the laboratory percentage of cocoons were found in vermiculite or shipped roundtrip from California to Massa- potting mix down to a depth of 2 cm (Table 3). chusetts did not significantly alter survivorship rates in comparison to the same treatments with- Harvesting of Franklinothrips orizabensis Pupae out icepacks (χ2 = 0.62, df = 1, p = 0.43). from Laboratory Colonies Temperature and humidity. Significant differ- ences in temperature (F = 2627, df = 3, 1448, p < Over a three day period, 48% of parafilm cones 0.0005) (Fig. 4A) and humidity (F = 173, df = 3, in cages were used by pupating F. orizabensis lar- 1448, p < 0.0005) (Fig. 4B) existed between poly-

TABLE 2. MEAN PERCENTAGE (± SE) OF FRANKLINOTHRIPS ORIZABENSIS IN EACH OF FOUR PUPATION MEDIA CON- TAINED IN PETRI DISHES THAT EITHER PUPATED SUCCESSFULLY AND EMERGED AS ADULTS, DIED AS LARVAE OR PUPAE, OR WERE NOT RECOVERED.

Pupation Media % Emerged % Dead % Unrecovered

Coarse vermiculite 80.00 ± 5.96 6.00 ± 3.06 14.00 ± 6.70 Fine vermiculite 0.00 ± 0.00 100 ± 0.00 0.00 ± 0.00 Parafilm cones 72.00 ± 4.42 8.00 ± 4.42 20.00 ± 5.16 Parafilm cones & fine vermiculite 66.67 ± 4.22 16.67 ± 6.15 16.67 ± 6.15 278 Florida Entomologist 84(2) June 2001

TABLE 3. MEAN PERCENTAGE (± SE) RECOVERY OF FRANKLINOTHRIPS ORIZABENSIS COCOONS FROM DIFFERENT LOCA- TIONS IN SMALL REARING CAGES.

Location of Cocoons in Cage % Recovered

Cage sleeves 10.40 ± 4.34 Cage corners 7.01 ± 1.49 Parafilm cones on potting media & cage floor 44.02 ± 6.86 Seed coat attached to cotyledons 9.61 ± 5.53 Potting soil 0-1cm deep 7.22 ± 3.52 1-2cm deep 3.89 ± 1.24 >2 cm deep 0.00 ± 0.00 Dead larvae & pupae 3.51 ± 1.24 Adults emerged into cages after inspection for cocoons 13.58 ± 3.10

styrene foam boxes that were retained in the lab- been inferred from analysis of leaf duff samples oratory. Boxes with ice-packs were significantly from commercial avocado orchards in southern cooler and had higher humidity rates, particu- California where F. orizabensis specimens have larly on the second day, presumably resulting been recovered to a depth of 2.5 cm into the soil from condensation on ice packs. The mean tem- (Hoddle et al. 1998). perature also increased in boxes with ice packs on Host plant abandonment rates by F. orizaben- the second day (Fig. 4A). Significant differences in sis larvae prior to pupation in avocado orchards temperature (F = 1422, df = 3, 2672, p < 0.0005) have not been quantified. Our estimates of larval (Fig. 4C) and humidity (F = 513, df = 3, 2672, p < abandonment on young avocado trees in the labo- 0.005) (Fig. 4D) existed between polystyrene foam ratory may have been over-estimated as young boxes that were shipped round trip from Califor- trees have fewer bark fissures that can be used as nia to Massachusetts. Boxes with ice-packs were pupation refugia in comparison to mature trees in significantly cooler, especially on the first day, and orchards. This suggestion is supported by higher had higher humidity rates, particularly on the recovery rates of F. orizabensis pupae on artificial second day, probably from condensation on ice branches with pupation refugia. In comparison to packs. Humidity in boxes without ice packs was phytophagous thrips, there is little published in- constant during the transit period (Fig. 4B). formation on pupation behavior and site selection by predatory thrips in field situations and this DISCUSSION area would benefit from more research. Coarse and fine vermiculite were unsuitable In the laboratory, the majority (82-99%) of re- pupation substrates for harvesting F. orizabensis covered late second instar F. orizabensis larvae pupae. In Petri dishes with coarse vermiculite, actively abandoned natural and artificial wooden pupating larvae constructed cocoons by attaching dowel branches to search for pupation sites below silk to walls and floors of dishes, and coarse ver- host plants. Predatory and phytophagous thrips miculite particles were then attached to cocoon have been captured emerging from the ground be- surfaces not adhered to the Petri dish. Pupating neath citrus trees (Childers et al. 1994), and pu- F. orizabensis could not be harvested in this me- pation beneath host plants is common for dia as cocoons were destroyed (i.e., cocoons were pestiferous phytophagous thrips (Grout et al. ripped open exposing pupae) when removal of 1986; Harrison 1963; Okada 1981; Reed & Rich vermiculite granules was attempted. Fine ver- 1975; Schweizer & Morse 1996, 1997; Tsuchida miculite was an unsuitable pupation substrate as 1997). Host plant abandonment in the field has it resulted in mortality of 100% of second instar

TABLE 4. UTILIZATION OF PARAFILM CONES BY FRANKLINOTHRIPS ORIZABENSIS IN MASS REARING CAGES.

Measured Variable Mean ± SE

Parafilm cones utilized by pupating larvae 48.33% ± 3.43 Male Franklinothrips orizabensis emerging from cones 39.49% ± 1.79 Female Franklinothrips orizabensis emerging from cones 60.51% ± 1.79 No. adult Franklinothrips orizabensis emerging per cone 1.31 ± 0.07 Pupal mortality in parafilm cones 2.35% ± 0.35 Hoddle et al.: Franklinothrips pupation biology 279

Fig. 3. Survivorship of aspirated adult male and female Franklinothrips orizabensis and pupae in parafilm cones either retained in the laboratory at ambient temperatures or cooled with ice packs, or shipped roundtrip from Riv- erside, California to Amherst, Massachusetts either with or without ice packs. Mean treatment proportions surviv- ing with the same letters are not significantly different (0.025 level of significance).

Fig. 4. Mean temperatures (A) and humidities (B) in polystyrene foam boxes with and without ice packs when retained in the laboratory for 48 h, and mean temperatures (C) and humidities (D) when polystyrene boxes were shipped round trip from Riverside, California to Amherst, Massachusetts with or without ice packs. Mean temper- atures and humidities followed by the same letters are not significantly different (0.05 level of significance). 280 Florida Entomologist 84(2) June 2001

F. orizabensis larvae. Placing parafilm cones in ACKNOWLEDGMENTS Petri dishes with fine vermiculite significantly in- creased larval and pupal survivorship. This work was supported in part by the Cali- Parafilm cones were utilized by late second in- fornia Avocado Commission. Dr. Roy Van Drie- star F. orizabensis larvae as pupation sites in sche, University of Massachusetts at Amherst Petri dishes, small rearing cages and laboratory kindly assisted with the shipping aspects of this colonies. In small rearing cages where known trial. We thank Dr. Miwa Takano-Lee for critically numbers of second instar larvae were deployed, evaluating this work. 44% of recovered F. orizabensis pupae were found in parafilm cones. Parafilm cones may have been desirable pupation sites because of the conical de- REFERENCES CITED sign (i.e., the tapering nature of the unit and its horizontal position) resulted in larvae being able ARAKAKI, N., AND S. OKAJIMA. 1998. Notes on the biol- to find a desirable wall width within the cone to ogy and morphology of a predatory thrips, Frankli- adhere silk strands during cocoon construction. In nothrips vespiformis (Crawford) (Thysanoptera: commercial mass rearing operations, the use of Aeolothripidae): First record from Japan. Entomol. parafilm cones or some other artificial pupation Sci. 1: 359-363. site under host plants (e.g., small, mass produced CHILDERS, C. C., S. NAKAHARA, AND R. J. BESHEAR. clear plastic cones) that are easily harvestable 1994. Relative abundance of bispinosa may be cost effective for rapidly collecting and and other species of Thysanoptera emerging from shipping F. orizabensis pupae. The use of elec- soil beneath “Navel” orange trees in Florida during spring flowering. J. Entomol. Sci. 29: 318-328. tronic scanning devices could conceivably assist in GRAFTON-CARDWELL, E. E., AND Y. OUYANG. 1995a. sorting cones with and without pupae as cones Augmentation of Euseius tularensis (Acari: Phy- with cocoons turn a blackish color as adults ma- toseiidae) in Citrus. Environ. Entomol. 24: 738-747. ture. This color change could aid electronic detec- GRAFTON-CARDWELL, E. E., AND Y. OUYANG. 1995b. Ma- tion and mechanical sorting (Petitt et al. 1996; nipulation of the predacious mite, Euseius tularensis Smittle et al. 1986; Whitten 1969; Wolf et al. (Acari: Phytoseiidae), with pruning for citrus thrips 1972). Alternatively, weight differences between control, pp. 251-254. In Thrips Biology and Manage- cones with and without pupae may be amenable ment. B. L. Parker, M. Skinner and T. Lewis (eds.). to air stream separation (Jackson et al. 1996). Me- Plenum Press, New York. GRAFTON-CARDWELL, E. 1997. Manipulation of E. tula- chanical harvesting and sorting of F. orizabensis rensis for improved thrips control. Citrograph 82: 8. pupae in artificial pupation media could signifi- GRAFTON-CARDWELL, E. E., Y. OUYANG, AND R. L. cantly reduce production costs for this predator. BUGG. 1999a. Leguminous cover crops to enhance F. orizabensis larvae that pupate in sites other population development of Euseius tularensis (Acari: than cones could be a source of adults for sustain- Phytoseiidae) in citrus. Biol. Control 16: 73-80. ing cage colonies and may mitigate the need to set GRAFTON-CARDWELL, E. E., Y. OUYANG, AND R. A. aside portions of the harvested product for colony STRIGGOW. 1999b. Predacious mites for control of cit- maintenance. It is possible that harvesting pupae rus thrips, Scirtothrips citri (Thysanoptera: Thripi- in cones and not returning individuals to colonies dae) in nursery crops. Biol. Control 14: 29-36. GROUT, T. G., J. G. MORSE, AND O. L. BRAWNER. 1986. that exhibit a preference for cones could select for Location of citrus thrips (Thysanoptera: Thripidae) individuals that do not utilize harvestable media pupation: tree or ground? J. Econ. Entomol. 79: 59-61. for pupation. GROUT, T. G., AND G. I. RICHARDS. 1990. The influence of Survivorship of F. orizabensis during shipping windbreak species on citrus thrips (Thysanoptera: was greatly enhanced if predators were shipped as Thripidae) populations and their damage to South pupae in parafilm cocoons. Adult F. orizabensis, es- African citrus orchards. J. Entomol. Soc. South Af- pecially males, were extremely sensitive to aspira- rica 53: 151-157. tion and transit stress, and high mortality resulted GROUT, T. G., AND G. I. RICHARDS. 1992a. The dietary ef- when adult males and females were shipped long fects of windbreak pollens on longevity and fecundity of a predacious mite Euseius addoensis addoensis distances. Transit mortality of adult and pupal F. (Acari: Phytoseiidae) found in citrus orchards in orizabensis was not significantly reduced by inclu- South Africa. Bull. Entomol. Res. 82: 317-320. sion of ice packs in polystyrene foam boxes. This GROUT, T. G., AND G. I. RICHARDS. 1992b. Euseius ad- may have been an artifact resulting from the time doensis addoensis, an effective predator of citrus of year this shipping trial was conducted as ship- thrips, Scirtothrips aurantii, in the eastern Cape ments were made during May when weather con- Province of South Africa. Experimental and Applied ditions were cool. Inclusion of ice packs with Acarology 15: 1-13. shipments of F. orizabensis pupae would be recom- HARRISON, J. O. 1963. Notes on the biology of banana mended, especially during summer. flower thrips, Frankliniella parvula, in the Domini- can Republic (Thysanoptera; Thripidae). Ann. Ento- Work is currently underway with commercial mol. Soc. America 56: 664-666. insectaries in California to implement mass rear- HODDLE, M. S., J. G. MORSE, P. PHILLIPS, AND B. FABER. ing of F. orizabensis for field trials against S. per- 1998. Progress on the management of avocado thrips. seae in avocado orchards. California Avocado Society Yearbook 82: 87-100. Hoddle et al.: Franklinothrips pupation biology 281

HODDLE, M. S., J. G. MORSE, P. PHILLIPS, B. FABER, Y. PARRELLA, M. P., AND T. LEWIS. 1997. Integrated Pest WEE, AND S. PEIRCE. 1999. Avocado thrips update. Management (IPM) in Field Crops, pp. 595-614. In Citrograph 84: 13-14. Thrips as Crop Pests. T. Lewis (ed.). CAB Interna- HODDLE, M. S., L. ROBINSON, K. DRESCHER, AND J. tional, Wallingford. JONES. 2000. Developmental and reproductive biol- PETITT, F. L., D. S. SCHUMAN, D. O. WIETLISBACH, AND ogy of a predatory Franklinothrips orizabensis (Thys- J. A. COFFELT. 1996. An automated system for collec- anoptera: Aeolothripidae). Biol. Control 18: 27-38. tion and counting of parasitized leafminer (Diptera: HODDLE, M. S. J. JONES, K. OISHI, D. MORGAN, AND L. ) larvae. Florida Entomol. 79: 450-454. ROBINSON. 2001. Evaluation of diets for the develop- REED, D. K., AND J. R. RICH. 1975. A new survey tech- ment and reproduction of Franklinothrips orizaben- nique for citrus thrips. J. Econ. Entomol. 68: 739-742. sis (Thysanoptera: Aeolothripidae). Bull. Entomol. SCHWEIZER, H., AND J. G. MORSE. 1996. Pupation sites Res. (in press). of Scirtothrips citri (Thysanoptera: Thripidae) and po- JACKSON C. G., H. T. CHAN JR., M. H. TANIGUCHI, D. B. tential management through increasing mortality of CHURCHILL, AND D. M. BILSLAND. 1996. Pneumatic instars on the ground. J. Econ. Entomol. 89: 1438-1445. air separation for the sorting of parasitized and un- SCHWEIZER, H., AND J. G. MORSE. 1997. Factors influ- parasitized fruit fly (Diptera; Tephritidae) puparia. encing survival of citrus thrips (Thysanoptera: J. Econ. Entomol. 89: 353-358. Thripidae) propupae and pupae on the ground. J. MCMURTRY, J. A., AND B. A. CROFT. 1997. Life-styles of Econ. Entomol. 90: 435-443. phytoseiid mites and their roles in biological control. SILVERS, C. 2000. Biological control of Scirtothrips per- Ann. Rev. of Entomol. 42: 291-321. seae Nakahara in California avocados: assessment of MORSE, J. G., T. S. BELLOWS JR., AND Y. IWATA. 1986. two generalist predators. MS Thesis, University of Technique for evaluating residual toxicity of pesti- California at Riverside. 103 pp. cides to motile . J. Econ. Entomol. 79: 281-283. SMITTLE, J., J. A. SEAWRIGHT, AND B. K. BIRKY. 1986. MUNGER, F. 1942. A method for rearing citrus thrips in Use of a commercial color sorter to separate sexes of the laboratory. J. Econ. Entomol. 35: 373-375. stable fly (Diptera: Muscidae) pupae from a genetic NAKAHARA, S. 1997. Scirtothrips perseae (Thysan- sexing strain. J. Econ. Entomol. 79: 877-878. optera: Thripidae), a new species infesting avocado TSUCHIDA, K. 1997. Adult emerging sites of Thrips in southern California. Insecta Mundi 11: 189-191. palmi (Thysanoptera; Thripidae) in an egg plant OKADA, T. 1981. Searching behavior for pupation place field with mulching sheet. Applied Entomol. and by late second instar larvae of Scirtothrips dorsalis Zool. 32: 246-249. Hood (Thysanoptera: Thripidae). Japanese Soc. Ap- WHITTEN, M. J. 1969. Automated sexing of pupae and plied Entomol. and Zool. 25: 10-16. its usefulness in control by sterile insects. J. Econ. PARKER, B. L., AND M. SKINNER. 1997. Integrated Pest Entomol. 62: 272-273. Management (IPM) in Tree Crops, pp. 615-638. In WOLF, W. W, M. W. STIMMAN, AND G. K. PARKER. 1972. Thrips as Crop Pests. T. Lewis (ed.). CAB Interna- A mechanism to aid sorting of male and female cab- tional, Wallingford. bage looper pupae. J. Econ. Entomol. 65: 1044-1047.